Adaptive data prefetch

Abstract

A system includes reception of a request to fetch a first dataset, the request including a first identifier associated with the first dataset, determination, in response to the request, of a first signature associated with the first dataset based on the first identifier and on respective identifiers of one or more datasets fetched consecutively with the first dataset, and determination, based on the determined first signature, of whether to update a first link associating a second dataset with the first dataset, the link indicating to prefetch the first dataset in response to a request to fetch the second dataset.

What is claimed is: 
     
       1. A system comprising:
 a storage system storing a plurality of database tables; 
 a buffer cache; and 
 a computer server to:
 receive a request to fetch a first dataset associated with one or more of the database tables, the request including a first identifier associated with the first dataset; 
 determine a first signature associated with the first dataset by application of a locality-sensitive hashing algorithm to the first identifier and respective identifiers of one or more datasets fetched contemporaneously with the first dataset; and 
 determine, by comparing a distance between the first signature and a second signature associated with a second dataset, whether to update a first link associating the second dataset with the first dataset, the link indicating that the first dataset should be prefetched and stored in the buffer cache in response to reception of a request to fetch the second dataset. 
 
 
     
     
       2. A system according to  claim 1 , the computer server further to:
 receive a request to fetch the second dataset; and 
 in response to the request to fetch the second dataset and based on the first link, prefetch the second dataset and store the second dataset in the buffer cache. 
 
     
     
       3. A system according to  claim 1 , wherein determination of whether to update the first link comprises determination of a second distance between the first signature and a third signature associated with an identifier of the one or more datasets, and
 comparison of the first distance and the second distance. 
 
     
     
       4. A system according to  claim 3 , wherein determination of whether to update the first link comprises determination of a third distance between the third signature and a fourth signature associated with a fourth dataset, and
 comparison of the third distance and the second distance. 
 
     
     
       5. A method comprising:
 receiving a request to fetch a first dataset, the request including a first identifier associated with the first dataset; 
 in response to the request, applying a locality-sensitive hashing algorithm to the first identifier and respective identifiers of one or more datasets fetched with the first dataset to determine a first signature associated with the first dataset; and 
 determining, by comparing a distance between the first signature and a second signature associated with a second dataset, whether to update a first link associating the second dataset with the first dataset, the link indicating to prefetch the first dataset in response to reception of a request to fetch the second dataset. 
 
     
     
       6. A method according to  claim 5 , further comprising:
 receiving a request to fetch the second dataset; and 
 in response to the request to fetch the second dataset and based on the first link, prefetching the second dataset and store the second dataset in a buffer cache. 
 
     
     
       7. A method according to  claim 5 , wherein determination of whether to update the first link comprises determination of a second distance between the first signature and a third signature associated with an identifier of the one or more datasets, and
 comparison of the first distance and the second distance. 
 
     
     
       8. A method according to  claim 7 , wherein determination of whether to update the first link comprises determination of a third distance between the third signature and a fourth signature associated with a fourth dataset, and
 comparison of the third distance and the second distance. 
 
     
     
       9. A non-transitory computer-readable medium storing program code executable by a processing unit of a computing system to:
 receive a request to fetch a first dataset, the request including a first identifier associated with the first dataset; 
 in response to the request, apply a locality-sensitive hashing algorithm to the first identifier and respective identifiers of one or more datasets fetched with the first dataset; and 
 determine, by comparing a distance between the first signature and a second signature associated with a second dataset, whether to update a first link associating the second dataset with the first dataset, the link indicating to prefetch the first dataset in response to a request to fetch the second dataset. 
 
     
     
       10. A medium according to  claim 9 , wherein determination of whether to update the first link comprises determination of a second distance between the first signature and a third signature associated with an identifier of the one or more datasets, and
 comparison of the first distance and the second distance. 
 
     
     
       11. A medium according to  claim 10 , wherein determination of whether to update the first link comprises determination of a third distance between the third signature and a fourth signature associated with a fourth dataset, and
 comparison of the third distance and the second distance.